Implement IJacute and ijacute for Dutch language

[arrowtype]

Relevant:

https://typedrawers.com/discussion/1294/how-do-you-implement-ijacute-and-ijacute

[arrowtype]

Current result of using combining accents:

@benkiel to fix with the addition of a mark feature, plus a little language feature

[arrowtype]

Note to self: once this is settled, also implement a locl feature for TRK (Turkish). It will probably be this:

  script latn;
    language TRK;
      sub i by idotaccent;
      sub i.mono by idotaccent.mono;
      sub i.italic by idotaccent.italic;
      sub i.simple by idotaccent.simple;

[arrowtype]

Testing this ... I think I’m close, but I might be able to shorten it with a few classes:

feature locl {
  script latn;
    language NLD; # Dutch
      lookup LigatureIJ {
        sub I acutecomb J acutecomb by IJacute;

        sub I J by IJ;

        sub Iacute J acutecomb by IJacute;

        sub i acutecomb j acutecomb by ijacute;
        sub i.mono acutecomb j acutecomb by ijacute;
        sub i.italic acutecomb j acutecomb by ijacute;
        sub i.italic acutecomb j.italic acutecomb by ijacute;
        sub i.simple acutecomb j acutecomb by ijacute;
        sub i.simple acutecomb j.italic acutecomb by ijacute;

        sub i j by ij;
        sub i.mono j by ij;
        sub i.italic j by ij;
        sub i.italic j.italic by ij;
        sub i.simple j by ij;
        sub i.simple j.italic by ij;

        sub iacute j acutecomb by ijacute;
        sub iacute.mono j acutecomb by ijacute;
        sub iacute.italic j acutecomb by ijacute;
        sub iacute.italic j.italic acutecomb by ijacute;
        sub iacute.simple j acutecomb by ijacute;
        sub iacute.simple j.italic acutecomb by ijacute;

      } LigatureIJ;
} locl;

[arrowtype]

Awesome – the following seems to work equally well as the code above, but is shorter and clearer.

@j_for_ij =      [j j.italic];
@i_for_ij =      [i      i.mono      i.italic      i.simple];
@iacute_for_ij = [iacute iacute.mono iacute.italic iacute.simple];

feature locl {
  script latn;
    language NLD; # Dutch
      lookup LigatureIJ {
        sub I acutecomb J acutecomb by IJacute;
        sub I J by IJ;
        sub Iacute J acutecomb by IJacute;

        sub @i_for_ij acutecomb @j_for_ij acutecomb by ijacute;
        sub @i_for_ij @j_for_ij by ij;
        sub @iacute_for_ij @j_for_ij acutecomb by ijacute;

      } LigatureIJ;
} locl;

[arrowtype]

Also useful: locl feature of Bungee fonts (view as "Raw" to properly see it):

https://github.com/djrrb/Bungee/blob/ce5aba00e2cd46a14577508306f8494c0f122271/sources/1-drawing/features/features.fea

[arrowtype]

WIP for this:

7cb331e

[aaronbell]

An alternate perspective :).

I was reviewing the Cascadia Code features the other day and was thinking about a couple of substitutions in the locl code—specifically the Catalan ldot and Dutch IJ substitutions.

It struck me that in both cases, I'm taking two code points and replacing them with a single glyph. While that makes sense in a proportional font for text environments, does it make sense in a coding font meant to operate in a monospace environment?

So I ended up removing the substitution for Ldot and ldot (the codepoints are still present though), and also changing the ij substitutions to just be:

	sub iacute j' by jacute;
	sub Iacute J' by Jacute;

It is a bit sad not to be able to use the nice accent IJ ligatures I made, but I do like the simplicity of it. I did leave the IJ and ij codepoints in for legacy purposes, but otherwise everything is just handled in monospace.

[benkiel]

@aaronbell that code will not cover everything, as there are dutch words (or loan words) that have an iacute but not a jacute.

[aaronbell]

@benkiel Oh, is that right? I've never seen that mentioned in all the articles about supporting the Dutch ijacute 😨 . From what I've been able to tell, it is pretty common to use code like I did above and that users would expect the j to be replaced with an accented j. Is that not the case?

Would you rather recommend only replacing the j only when followed by a combining acute?

[benkiel]

Yes, that is safer. This comes from a discussion with EvB who pointed out that the IJacute has edge cases

[aaronbell]

@benkiel Thanks! Well, if that's the case, then maybe I'll get rid of those substitutions entirely as the ccmp will take care of it. (sub [i j i-cy je-cy]' @CombiningTopAccents by [idotless jdotless idotless jdotless];)